On-street parking places parallel to the curb have variable lengths, depending on the size of the car that emptied the place. Thus, a Smart Parking system should publish such places only to drivers whose car is shorter than the available parking length. We developed a crowdsourced Smart Parking app, based on implicit interaction, intending to publish an available parking spot only to drivers whose car can fit the existing place. This app detects the type of parking (parallel vs angle or perpendicular) using machine learning on the driver’s smartphone and considers the length of the cars involved.